Real-time computing

Computing in which correctness depends on producing results within specified time constraints as well as producing correct results. Missing a deadline can make an otherwise correct result useless or unsafe.

Real-time operating system: An operating system designed to provide predictable response times for tasks with timing constraints. It schedules tasks to make response times predictable.
Deadline (computing): The latest time by which a computing task must complete to satisfy its specification. A deadline defines the central timing obligation of a real-time task.
Embedded system: A computer system built into a larger device to perform dedicated functions. Many embedded devices must respond to physical events on time.
Batch processing: The execution of accumulated jobs in groups, usually without immediate interaction with each job's input. Batch workloads prioritize throughput rather than fixed response deadlines.
Real-time scheduling: The assignment of processor time to tasks with deadlines or other timing constraints. Its policies decide which task runs when deadlines compete.
Latency (engineering): The delay between an input, event, or request and the corresponding response or output. Latency consumes the time available before a deadline.
Industrial control system: A computer-based system that monitors or controls industrial processes and equipment. Late control actions can disrupt processes or damage equipment.
Best-effort delivery: A service model that attempts delivery without guaranteeing timing, reliability, or resource availability. Real-time guarantees go beyond best-effort responses.
Worst-case execution time: An upper bound on the time a program or task can take to execute under specified conditions. Deadline guarantees require bounds on task execution time.
Jitter: Variation in the timing of repeated events, signals, or task responses. Unpredictable timing variation can break otherwise adequate schedules.

Linked from 10 pages